►
Description
Estuary - presented by @jlogelin at IPFS bing 2022 - IPFS Implementations - https://2022.ipfs-thing.io
A
A
A
Initially
the
data
is
pinned
onto
the
estuary
ipfs
node,
the
estuary
node
is
an
ipfs
node
for
hot
storage.
So
as
soon
as
you
put
it
up
there,
it's
available
right
away,
and
then
this
process
is
kind
of
replicated
or
the
process
of
pushing
the
data
out
onto
the
filecoin
network
is
replicated
six
times
to
ensure
the
data
is
always
accessible.
A
So
yeah
this
is
me
web
2,
trad
tech
guy.
Coming
in
saying
why?
Why
don't
I
just
write
it
myself?
I
can
just
look
at
the
source
code
and
build
it
myself,
yeah.
No,
not
so
much
estuary
uses
several
libraries
and
supporting
protocols
in
an
attempt
to
amalgamate.
I
hope
I
get
this
right.
Lotus,
kubo
and
friends.
A
Does
that
sound,
ish,
right,
yeah,
okay,
I'll
go
with
that,
so
it
makes
the
ipfs
node
pretty
complicated
with
a
lot
of
moving
parts,
especially
if
you're
a
guy
like
me,
that's
been
in
sort
of
the
traditional
tech
world
and
yeah
digging
into
the
code.
This
is
kind
of
what
I
have
to
deal
with,
where
what
one
would
have
to
deal
with
if
you
wanted
to
to
sort
of
credit
yourself.
So
the
mission
is
ease
of
use
and
I
got
a
quick
little
video
here
to
show.
A
A
Up
and
so
once
you
push
it
up,
you
see
you
can
see
the
sort
of
six
box
indicator
to
indicate
okay.
These
are
the
receipts
that
that
came
back
from
the
file
going
network
that
this
can
take
a
little
bit
of
time
to
happen.
But
again,
you've
got
the
the
hot
storage
and
the
sort
of
cold
storage
combination,
also
extremely
important
to
have
accessibility
to
your
files.
All
around
the
world
estuary
makes
this
a
quote-unquote
five-minute
problem
right.
You
guys
just
saw
it.
A
A
A
So
estuary
has
two
primary
components:
the
prime
the
main
node
or
primary
node,
as
it's
called,
is
like
that
big
giant
box
in
the
middle,
and
then
it's
got
shuttles
as
well.
That
helps
scale
the
pinning
of
data,
the
no
the
node
all
right.
I
guess
this
talking
about
the
data
flow,
the
clients
communicate
with
the
the
estuary
nodes,
the
primary
node
or
the
the
secondary
shuttle
nodes.
They
can.
You
can
interact
with
those
directly
as
well
to
push
your
data
up.
A
Node,
checks
for
registered
service
providers
create
six
deals
with
the
six
x
replication
and
then
transfers
it
over
using
the
bitsoft
bitswap
protocol
and
we
utilize
a
local
client
called
fill
client,
which
is
like
a
lightweight
filecoin
client
maintained
by
application.
Research,
which
is
the
group
I
work
for
the
primary
node
itself,
has
seven
components,
probably
the
most
important
or
like
the
secret
sauce
is
the
content
manager.
A
This
is
where
the
verification
and
happens
with
the
the
6x
copies
across
the
falcon
network.
Then
there's
like
the
deal
maker,
which
is
like
a
storage
provider
check
when
the
storage
provider
is
added
to
the
system
and
it
kind
of
spreads
the
deals
like
50
50
between
top
tier
and
lower
tier.
So
it
does
a
balance.
A
You
could
potentially
download
the
node
yourself
the
code
yourself
and
adjust
that
that's
that's
just
what
we've
chose
and
importantly
the
staging
manager
writes.
Two
ipfs
are
right
to
the
manages
the
ipfs
pinning
which,
as
of
I
think
last
month,
is
100
compatible
with
the
ipfs
pinning
api,
which
is
pretty
rad.
A
A
So
yeah,
like
I
said
before
shuttles,
are,
are
how
we
horizontally
distribute
payloads
into
estuary.
The
primary
node
can
instruct
shuttles
to
talk
to
each
other
ipfs
nodes,
shuttles
are
actually
shuttles
themselves
and
sorry.
Shuttles
are
full
ipfs
nodes
rather
and
shuttles
can
actually
be
co-located
with
the
data,
which
is
kind
of
handy
if
you're
pushing
a
lot
of
data
up
to
certain
areas
of
the
globe.
A
And
sort
of
the
other
piece
I
guess
to
estuary
is
okay:
we've
got
this
kind
of
cool
broker,
that's
up
and
running.
How
do
we
actually
push
data
up
so
there's
there's
a
small
project,
that's
sort
of
kind
of
sitting
on
its
own.
Now
it
used
to
be
part
of
the
primary
project
we
just
kind
of
moved
it
out
on
its
own
and
it's
called
barge.
So
it's
a
cli
tool
that
directly
uses
estuary
api
to
stream
and
upload
files
into
the
filecoin
network.
A
So
how
do
web
2
or
web
developers
actually
use
this
or
how
to
what
developers
actually
use
this
again
with
the
feature
of?
Is
it's
ease
of
use
right?
So
this
this
page?
If
you
guys
go
to
docs.estuary.tech,
it'll
kind
of
show
you
this
this
docs
page
and
what's
kind
of
cool,
is
cake
built.
This
really
neat
sort
of
interactive
way
to
sort
of
play
with
the
different
api
calls
to
a
react
component
and
it's
live.
A
A
A
And
then,
of
course,
we
have
access
to
deeper
examples
into
the
apis,
such
as
pinning
deals
minor,
ap
minor
apis.
You
can
actually
like
direct
your
directed
to
to
any
storage
provider
that
you
want
supported
by
the
the
node
and
then
there's
an
admin
system
api
as
well,
and
so
again
I
I
keep
harping
on
it
but
kind
of
as
a
new
guy
coming
in.
I
found
this
to
be
really
handy.
A
A
A
A
A
A
A
A
Broaden
our
network
and,
most
importantly,
I
think
in
the
short
term,
is
to
get
auto
retrieve
up
and
running,
and
so
that's
the
team
cake
has
recently
changed
his
name
in
case
anybody
didn't
know
it's
now
that
and
I
am
over
and
it's
too
bad,
because
I
wanted
to
talk
about
this.
But
if
anybody
wants
to
talk
about
the
interplanetary
container
registry
with
me,
I've
got
like
four
or
five
slides
that
I
just
like
threw
together
right
now,
but
thanks
a
lot.